Open AccessA simplified machine learning approach for recognizing human activity
Author(s) -
Abdul Lateef Haroon P S,
U. Eranna
Publication year - 2019
Publication title -
international journal of power electronics and drive systems/international journal of electrical and computer engineering
Language(s) - English
Resource type - Journals
eISSN - 2722-2578
pISSN - 2722-256X
DOI - 10.11591/ijece.v9i5.pp3465-3473
Subject(s) - computer science , boosting (machine learning) , artificial intelligence , activity recognition , machine learning , process (computing) , event (particle physics) , scope (computer science) , feature extraction , pattern recognition (psychology) , real time computing , computer vision , physics , quantum mechanics , programming language , operating system
With the wide ranges of real-time event feed capturing devices, there has been significant progress in the area of digital image processing towards activity detection and recognition. Irrespective of the presence of various such devices, they are not adequate to meet dynamic monitoring demands of the visual surveillance system, and their features are highly limited towards complex human activity recognition system. Review of existing system confirms that still there is a large scope of enhancement as they lack applicability to real-life events and also doesn't offer optimal system performance. Therefore, the proposed manuscript presents a model for activity recognition system where the accuracy of recognition operation and system performance are retained with good balance. The study presents a simplified feature extraction process from spatial and temporal traits of the event feeds that is further subjected to the machine learning mechanism for boosting recognition performance